Starter prompts
4 ways to start with SalesData.
Map columns
→ Fuzzy schema
▸ Preview prompt
Build a fuzzy column mapper that handles Excel sheets with revenue/sales/total_sales variants and reports unmapped columns.
Pipeline
→ End to end
▸ Preview prompt
Design a pipeline that ingests weekly sales Excels into Postgres with MTD/YTD/Year End normalization.
Audit
→ Prove imports
▸ Preview prompt
Add an audit trail so every row in our metrics table traces back to a source file, row, and import timestamp.
Errors
→ Quarantine plan
▸ Preview prompt
Quarantine rows that fail validation, notify the right rep manager, and reprocess after manual fix without duplicating data.
What it does
Tasks SalesData ships every week.
Ingestion
- Watch + detect new files
- Fuzzy column mapping
- Currency + format cleanup
- Rep matching + dedupe
Loading
- Atomic bulk inserts
- Audit trail + source file
- Quarantine bad rows
- Idempotent re-runs
Worked sample
A real SalesData chat.
Pairs well with